Table 4.

Multivariate Association of Carotid Intima–Media Thickness with Various Risk Factors.^*

Risk Factor	Common Carotid Intima–Media Thickness			Combined Intima–Media Thickness
		Semipartial R2	P Value		Semipartial R2	P Value
	Estimate of β Coefficient ±SE	%		Estimate of β Coefficient ±SE	%
Age (per year of age)	0.0029±0.0006	1.32	<0.001	0.0248±0.0071	0.49	<0.001
Sex (female vs. male)	−0.0173±0.0055	0.50	0.002	−0.2506±0.0689	0.53	<0.001
Year 1 intima–media thickness	0.6188±0.0361	14.79	<0.001	0.5961±0.0238	24.90	<0.001
Smoking (yes vs. no)	0.0265±0.0068	0.77	<0.001	0.3737±0.0851	0.77	<0.001
Systolic blood pressure (per mm Hg)	0.0008±0.0002	0.61	<0.001	0.0092±0.0029	0.42	0.001
Ultrasonography equipment used (12 combinations)†	—	1.64	<0.001	—	1.92	<0.001
Treatment group (conventional vs. intensive)	−0.0453±0.0274	0.14	0.10	−0.5433±0.3435	0.10	0.12
Interaction between age and treatment group	0.0017±0.0008	0.24	0.03	0.0201±0.0096	0.18	0.04
Overall treatment effect (treatment effect as a function of age)‡	§	0.57 0.40	0.004	‖	0.42 0.52	0.005
Total R2 (%)	40			52

^*The multiple regression model used the year 6 intima–media thickness as the outcome and fit the listed covariates simultaneously.

^†Eleven estimates are omitted.

^‡The overall treatment effect was determined by two regression coefficients: treatment group and the interaction between age and treatment group. The use of either coefficient alone did not reflect the overall treatment effect. The overall treatment effect was based on the general linear test with 2 degrees of freedom for the numerator.

^§For the common carotid artery, β = (−0.0453) + (0.0017) × age, and the standard error = (7.5×10⁻⁴) − (4.1×10⁻⁵) × age + (6.4×10⁻⁷) × age².

^‖For the combined value, β = (−0.5433) + (0.0201) × age, and the standard error = (0.1180) − (0.0065) × age + (9.2×10⁻⁵) × age².